From: Pranjal Shrivastava <praan@google.com>
To: Nicolin Chen <nicolinc@nvidia.com>
Cc: jgg@nvidia.com, kevin.tian@intel.com, will@kernel.org,
robin.murphy@arm.com, joro@8bytes.org, yi.l.liu@intel.com,
peterz@infradead.org, jsnitsel@redhat.com,
linux-arm-kernel@lists.infradead.org, iommu@lists.linux.dev,
linux-kernel@vger.kernel.org, patches@lists.linux.dev,
baolu.lu@linux.intel.com
Subject: Re: [PATCH v2 10/14] iommu/arm-smmu-v3: Replace arm_vsmmu_alloc with arm_vsmmu_init
Date: Mon, 16 Jun 2025 22:43:22 +0000 [thread overview]
Message-ID: <aFCeCvKkm1MvT3xP@google.com> (raw)
In-Reply-To: <64e4b4c33acd26e1bd676e077be80e00fb63f17c.1749882255.git.nicolinc@nvidia.com>
On Fri, Jun 13, 2025 at 11:35:22PM -0700, Nicolin Chen wrote:
> To ease the for-driver iommufd APIs, get_viommu_size and viommu_init ops
> are introduced.
>
> Sanitize the inputs and report the size of struct arm_vsmmu on success, in
> arm_smmu_get_viommu_size().
>
> Place the type sanity at the last, becase there will be soon an impl level
> get_viommu_size op, which will require the same sanity tests prior. It can
> simply insert a piece of code in front of the IOMMU_VIOMMU_TYPE_ARM_SMMUV3
> sanity.
>
That's what I was wondering, so we plan to replace the impl->vsmmu_alloc
op as well?
> The core will ensure the viommu_type is set to the core vIOMMU object, and
> pass in the same dev pointer, so arm_vsmmu_init() won't need to repeat the
> same sanity tests but to simply init the arm_vsmmu struct.
>
> Remove the arm_vsmmu_alloc, completing the replacement.
>
> Signed-off-by: Nicolin Chen <nicolinc@nvidia.com>
> ---
> drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.h | 11 +++--
> .../arm/arm-smmu-v3/arm-smmu-v3-iommufd.c | 46 ++++++++++---------
> drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.c | 3 +-
> 3 files changed, 32 insertions(+), 28 deletions(-)
>
> diff --git a/drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.h b/drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.h
> index ea41d790463e..bb39af84e6b0 100644
> --- a/drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.h
> +++ b/drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.h
> @@ -1034,18 +1034,19 @@ struct arm_vsmmu {
>
> #if IS_ENABLED(CONFIG_ARM_SMMU_V3_IOMMUFD)
> void *arm_smmu_hw_info(struct device *dev, u32 *length, u32 *type);
> -struct iommufd_viommu *arm_vsmmu_alloc(struct device *dev,
> - struct iommu_domain *parent,
> - struct iommufd_ctx *ictx,
> - unsigned int viommu_type);
> +size_t arm_smmu_get_viommu_size(struct device *dev,
> + enum iommu_viommu_type viommu_type);
> +int arm_vsmmu_init(struct iommufd_viommu *viommu,
> + struct iommu_domain *parent_domain);
> int arm_smmu_attach_prepare_vmaster(struct arm_smmu_attach_state *state,
> struct arm_smmu_nested_domain *nested_domain);
> void arm_smmu_attach_commit_vmaster(struct arm_smmu_attach_state *state);
> void arm_smmu_master_clear_vmaster(struct arm_smmu_master *master);
> int arm_vmaster_report_event(struct arm_smmu_vmaster *vmaster, u64 *evt);
> #else
> +#define arm_smmu_get_viommu_size NULL
> #define arm_smmu_hw_info NULL
> -#define arm_vsmmu_alloc NULL
> +#define arm_vsmmu_init NULL
>
> static inline int
> arm_smmu_attach_prepare_vmaster(struct arm_smmu_attach_state *state,
> diff --git a/drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3-iommufd.c b/drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3-iommufd.c
> index e4fd8d522af8..9f59c95a254c 100644
> --- a/drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3-iommufd.c
> +++ b/drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3-iommufd.c
> @@ -382,25 +382,14 @@ static const struct iommufd_viommu_ops arm_vsmmu_ops = {
> .cache_invalidate = arm_vsmmu_cache_invalidate,
> };
>
> -struct iommufd_viommu *arm_vsmmu_alloc(struct device *dev,
> - struct iommu_domain *parent,
> - struct iommufd_ctx *ictx,
> - unsigned int viommu_type)
> +size_t arm_smmu_get_viommu_size(struct device *dev,
> + enum iommu_viommu_type viommu_type)
> {
> - struct arm_smmu_device *smmu =
> - iommu_get_iommu_dev(dev, struct arm_smmu_device, iommu);
> struct arm_smmu_master *master = dev_iommu_priv_get(dev);
> - struct arm_smmu_domain *s2_parent = to_smmu_domain(parent);
> - struct arm_vsmmu *vsmmu;
> -
> - if (viommu_type != IOMMU_VIOMMU_TYPE_ARM_SMMUV3)
> - return ERR_PTR(-EOPNOTSUPP);
> + struct arm_smmu_device *smmu = master->smmu;
>
> if (!(smmu->features & ARM_SMMU_FEAT_NESTING))
> - return ERR_PTR(-EOPNOTSUPP);
> -
> - if (s2_parent->smmu != master->smmu)
> - return ERR_PTR(-EINVAL);
> + return 0;
>
> /*
> * FORCE_SYNC is not set with FEAT_NESTING. Some study of the exact HW
> @@ -408,7 +397,7 @@ struct iommufd_viommu *arm_vsmmu_alloc(struct device *dev,
> * any change to remove this.
> */
> if (WARN_ON(smmu->options & ARM_SMMU_OPT_CMDQ_FORCE_SYNC))
> - return ERR_PTR(-EOPNOTSUPP);
> + return 0;
>
> /*
> * Must support some way to prevent the VM from bypassing the cache
> @@ -420,19 +409,32 @@ struct iommufd_viommu *arm_vsmmu_alloc(struct device *dev,
> */
> if (!arm_smmu_master_canwbs(master) &&
> !(smmu->features & ARM_SMMU_FEAT_S2FWB))
> - return ERR_PTR(-EOPNOTSUPP);
> + return 0;
>
> - vsmmu = iommufd_viommu_alloc(ictx, struct arm_vsmmu, core,
> - &arm_vsmmu_ops);
> - if (IS_ERR(vsmmu))
> - return ERR_CAST(vsmmu);
> + if (viommu_type != IOMMU_VIOMMU_TYPE_ARM_SMMUV3)
> + return 0;
> +
> + return VIOMMU_STRUCT_SIZE(struct arm_vsmmu, core);
> +}
> +
> +int arm_vsmmu_init(struct iommufd_viommu *viommu,
> + struct iommu_domain *parent_domain)
> +{
> + struct arm_vsmmu *vsmmu = container_of(viommu, struct arm_vsmmu, core);
> + struct arm_smmu_device *smmu =
> + container_of(viommu->iommu_dev, struct arm_smmu_device, iommu);
> + struct arm_smmu_domain *s2_parent = to_smmu_domain(parent_domain);
> +
> + if (s2_parent->smmu != smmu)
> + return -EINVAL;
>
> vsmmu->smmu = smmu;
> vsmmu->s2_parent = s2_parent;
> /* FIXME Move VMID allocation from the S2 domain allocation to here */
> vsmmu->vmid = s2_parent->s2_cfg.vmid;
>
> - return &vsmmu->core;
> + viommu->ops = &arm_vsmmu_ops;
> + return 0;
> }
Seems much better now that the driver doesn't need to callback to the
core for allocating viommu. One quick question though I see we've
removed the following too:
if (master->smmu->impl_ops &&master->smmu->impl_ops->vsmmu_alloc)
vsmmu = master->smmu->impl_ops->vsmmu_alloc(
master->smmu, s2_parent,
ictx, viommu_type,
user_data);
Not sure why don't I see that in the diffs.. do we plan to split this
into an impl-specific size and init too?
>
> int arm_vmaster_report_event(struct arm_smmu_vmaster *vmaster, u64 *evt)
> diff --git a/drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.c b/drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.c
> index 10cc6dc26b7b..181d07bc1a9d 100644
> --- a/drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.c
> +++ b/drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.c
> @@ -3688,7 +3688,8 @@ static struct iommu_ops arm_smmu_ops = {
> .get_resv_regions = arm_smmu_get_resv_regions,
> .page_response = arm_smmu_page_response,
> .def_domain_type = arm_smmu_def_domain_type,
> - .viommu_alloc = arm_vsmmu_alloc,
> + .get_viommu_size = arm_smmu_get_viommu_size,
> + .viommu_init = arm_vsmmu_init,
> .user_pasid_table = 1,
> .pgsize_bitmap = -1UL, /* Restricted during device attach */
> .owner = THIS_MODULE,
> --
> 2.43.0
>
With that,
Reviewed-by: Pranjal Shrivastava <praan@google.com>
next prev parent reply other threads:[~2025-06-16 22:47 UTC|newest]
Thread overview: 52+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-06-14 6:35 [PATCH v2 00/14] iommufd: Prepare for IOMMUFD_OBJ_HW_QUEUE Nicolin Chen
2025-06-14 6:35 ` [PATCH v2 01/14] iommufd: Apply obvious cosmetic fixes Nicolin Chen
2025-06-16 3:23 ` Baolu Lu
2025-06-14 6:35 ` [PATCH v2 02/14] iommufd: Drop unused ictx in struct iommufd_vdevice Nicolin Chen
2025-06-16 3:23 ` Baolu Lu
2025-06-14 6:35 ` [PATCH v2 03/14] iommufd: Use enum iommu_viommu_type for type in struct iommufd_viommu Nicolin Chen
2025-06-16 3:24 ` Baolu Lu
2025-06-16 21:45 ` Pranjal Shrivastava
2025-06-14 6:35 ` [PATCH v2 04/14] iommufd: Use enum iommu_veventq_type for type in struct iommufd_veventq Nicolin Chen
2025-06-16 3:24 ` Baolu Lu
2025-06-16 21:47 ` Pranjal Shrivastava
2025-06-14 6:35 ` [PATCH v2 05/14] iommufd: Return EOPNOTSUPP for failures due to driver bugs Nicolin Chen
2025-06-16 3:25 ` Baolu Lu
2025-06-16 12:28 ` Jason Gunthorpe
2025-06-16 21:49 ` Pranjal Shrivastava
2025-06-19 5:40 ` Tian, Kevin
2025-06-14 6:35 ` [PATCH v2 06/14] iommu: Introduce get_viommu_size and viommu_init ops Nicolin Chen
2025-06-16 3:25 ` Baolu Lu
2025-06-16 12:43 ` Jason Gunthorpe
2025-06-16 22:10 ` Pranjal Shrivastava
2025-06-19 5:42 ` Tian, Kevin
2025-06-14 6:35 ` [PATCH v2 07/14] iommufd/viommu: Support " Nicolin Chen
2025-06-16 3:26 ` Baolu Lu
2025-06-16 12:45 ` Jason Gunthorpe
2025-06-19 5:45 ` Tian, Kevin
2025-06-14 6:35 ` [PATCH v2 08/14] iommufd/selftest: Drop parent domain from mock_iommu_domain_nested Nicolin Chen
2025-06-16 12:46 ` Jason Gunthorpe
2025-06-19 5:46 ` Tian, Kevin
2025-06-14 6:35 ` [PATCH v2 09/14] iommufd/selftest: Replace mock_viommu_alloc with mock_viommu_init Nicolin Chen
2025-06-16 12:46 ` Jason Gunthorpe
2025-06-14 6:35 ` [PATCH v2 10/14] iommu/arm-smmu-v3: Replace arm_vsmmu_alloc with arm_vsmmu_init Nicolin Chen
2025-06-16 10:03 ` Will Deacon
2025-06-16 12:49 ` Jason Gunthorpe
2025-06-16 22:43 ` Pranjal Shrivastava [this message]
2025-06-17 2:15 ` Nicolin Chen
2025-06-17 5:43 ` Pranjal Shrivastava
2025-06-14 6:35 ` [PATCH v2 11/14] iommu: Deprecate viommu_alloc op Nicolin Chen
2025-06-16 3:26 ` Baolu Lu
2025-06-16 12:49 ` Jason Gunthorpe
2025-06-16 22:46 ` Pranjal Shrivastava
2025-06-14 6:35 ` [PATCH v2 12/14] iommufd: Move _iommufd_object_alloc out of driver.c Nicolin Chen
2025-06-16 3:26 ` Baolu Lu
2025-06-14 6:35 ` [PATCH v2 13/14] iommufd: Introduce iommufd_object_alloc_ucmd helper Nicolin Chen
2025-06-16 3:27 ` Baolu Lu
2025-06-16 22:52 ` [PATCH v2 13/14] iommufd: Introduce iommufd_object_alloc_ucmd helpery Pranjal Shrivastava
2025-07-09 5:31 ` [PATCH v2 13/14] iommufd: Introduce iommufd_object_alloc_ucmd helper Xu Yilun
2025-07-10 5:32 ` Tian, Kevin
2025-07-10 18:21 ` Nicolin Chen
2025-06-14 6:35 ` [PATCH v2 14/14] iommufd: Apply the new " Nicolin Chen
2025-06-16 3:27 ` Baolu Lu
2025-06-19 5:49 ` Tian, Kevin
2025-06-19 18:46 ` [PATCH v2 00/14] iommufd: Prepare for IOMMUFD_OBJ_HW_QUEUE Jason Gunthorpe
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=aFCeCvKkm1MvT3xP@google.com \
--to=praan@google.com \
--cc=baolu.lu@linux.intel.com \
--cc=iommu@lists.linux.dev \
--cc=jgg@nvidia.com \
--cc=joro@8bytes.org \
--cc=jsnitsel@redhat.com \
--cc=kevin.tian@intel.com \
--cc=linux-arm-kernel@lists.infradead.org \
--cc=linux-kernel@vger.kernel.org \
--cc=nicolinc@nvidia.com \
--cc=patches@lists.linux.dev \
--cc=peterz@infradead.org \
--cc=robin.murphy@arm.com \
--cc=will@kernel.org \
--cc=yi.l.liu@intel.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).